Reitse Eskens does a performance comparison:
A few days ago, I heard someone stating that Group By was much quicker than Distinct. Less disk impact, less memory etc.
So, I thought I’d find out if it’s true or not because I found it interesting. I always thought there was no difference. I tested a single small table and found no difference in speed, reads or execution plan. But that’s no real world example. Usually the tables contain a lot of data and are joined to other tables.
Click through for the results of Reitse’s analysis.